Effect of arsenic on the risk of gestational diabetes mellitus: a systematic review and meta-analysis

This meta-analysis aimed to summarize the main studies on the correlation between As and GDM incidence, revealing the relationship between As exposure and GDM. However, considering the heterogeneity of merged studies, a detailed explanation of the results is crucial. In addition, sensitivity analysis revealed that despite significant heterogeneity, the results were statistically significant and robust. Subgroup analyses indicated that heterogeneity could be attributed to the year of publication.

As an increasing number of young women are diagnosed with hyperglycemia or overt diabetes, people are becoming increasingly worried about GDM. It is caused by several risk factors, including obesity, previous GDM history, and familial history of type 2 diabetes [40]. The pathophysiology of GDM is multifactorial and involves both genetic predispositions and environmental triggers that contribute to its development. A key factor in the pathogenesis of GDM is insulin resistance [41], which is exacerbated during pregnancy due to hormonal changes, leading to inadequate insulin compensation and hyperglycemia.

Recent studies have highlighted the role of assisted reproductive technology (ART) as a significant variable in assessing GDM risk. Women who conceive through ART may have a predisposition to metabolic disorders, including GDM, due to underlying factors such as polycystic ovary syndrome (PCOS) and endometriosis, which are often associated with infertility treated by ART. Furthermore, the hormonal treatments involved in ART, such as ovulation induction, can alter glucose metabolism, increasing the risk of developing GDM. The influence of ART on GDM risk underscores the importance of considering the mode of conception in the assessment and management of GDM. For instance, a meta-analysis [42] stratified by mode of conception, disease location, and severity revealed a greater risk of GDM in women with endometriosis, especially those who conceived through ART, than in those conceived through natural conception. These findings suggest that ART may compound the risk of GDM in women already predisposed to pregnancy due to underlying reproductive disorders.

Moreover, demographic and clinical data play a crucial role in understanding the risk factors and mechanisms underlying GDM. Incorporating detailed demographic and clinical data, including prepregnancy weight, age, family history of diabetes, and previous GDM history, can provide insights into the individual risk profiles of pregnant women. A comprehensive approach that includes these variables, as detailed in recent research [43], can enhance our understanding of the multifaceted etiology of GDM and improve risk stratification and management strategies.

Building on this foundation, the variability of GDM diagnostic criteria poses certain challenges. The criteria proposed by the International Association of Diabetes and Pregnancy Study Groups (IADPSG) have been widely advocated, yet significant disagreements exist in the established diagnostic standards for GDM [44, 45]. These criteria are based on the 75-g oral glucose tolerance test (OGTT), which diagnoses GDM when fasting blood glucose is ≥ 5.1 mmol/L, 1-hour blood glucose is ≥ 10.0 mmol/L, or 2-hour blood glucose is ≥ 8.0 mmol/L. Some approved standards and guidelines exist, such as those from the World Health Organization (WHO) and the Australian Diabetes Association (ADIPS). Blood glucose thresholds differ at different times [46]. This disagreement highlights the various interpretations and implementations of GDM diagnosis in different regions. Throughout pregnancy, changes in glucose regulation are crucial for ensuring adequate nutrition for the fetus. Existing studies using high-insulin normoglycemic clamps have shown a 56% decrease in insulin sensitivity and a 30% increase in basal endogenous glucose production during the advanced stages of pregnancy. To counterbalance these changes, pancreatic β-cells enhance insulin secretion to maintain blood glucose stability [47]. Further studies suggested that any insulin resistance correlated with a normal pregnancy will rapidly reverse after delivery, indicating the regulatory role of placental factors in these physiological changes. A link between the environmental pollutant As and GDM has been established, as As can penetrate the placental barrier and affect glucose metabolism in both mothers and fetuses [2].

The universal correlation of As with numerous pathological conditions has attracted global concern, as As affects countries at all levels of development. Despite the severe health risks associated with As exposure, contact is often unintentional, primarily because of the consumption of contaminated water and food once inside the body. It induces metabolic disorders through multiple pathways and causes inflammatory responses, triggering oxidative stress and inflammation [48]. Moreover, As impairs glucose tolerance predominantly through the dysfunction of insulin-secreting β-cells rather than through an increase in peripheral insulin resistance [49]. The liver can use the methyl group in S-adenosylmethionine (SAM) to methylate inorganic As, forming monomethyl arsenic acid (MMA) and dimethyl arsenic acid (DMA), both of which are organic As with reduced toxicity [50]. Therefore, a higher efficiency of methylation results in lower toxicity due to increased conversion rates. Understanding the differences in toxicity between organic and inorganic As is crucial. Some inorganic As is metabolized into o-arsenate (As³⁺), a highly toxic metabolite linked to As poisoning. The metabolized As are primarily excreted through the kidneys. Some are excreted through the urine, and others are excreted through the bile. Strict mitigation measures are needed, as they can penetrate various biological barriers and affect multiple organs and systems.

Previous observational studies have suggested an indirect association between GDM and As exposure [17‐20, 30, 31, 33]. Conversely, recent studies suggest that this relationship lacks statistical significance [22‐25]. Given these different findings, we performed this meta-analysis of the latest evidence to explain the degree of connection between GDM and As exposure. The studies published between 2021 and 2023, cited by Chen et al. [22], had no statistically significant associations. They concluded that there was no significant association between total inorganic As and GDM (OR 0.77, 95% CI: 0.33 to 1.79). Similarly, Jia et al. [23] measured As levels in hair and found no substantial association between As and GDM (OR 0.91, 95% CI: 0.77 to 1.07). Furthermore, Zhang et al. [24] studied the combined impact of As and one-carbon metabolism (OCM) on GDM using urine. They found no significant association between total As and GDM (OR 0.90, 95% CI: 0.44 to 1.82). Another recent study in China by Liang et al. [25] used urine measurements and found no significant association between inorganic As and GDM (OR 0.66, 95% CI: 0.30 to 1.45). Considering the inconsistent results in the literature, this emphasizes the importance of ongoing investigations and rigorous analyses.

In the preliminary meta-analysis of As and GDM, nine studies involving 1,984 GDM patients were merged [51]. The final result (OR 1.56, 95% CI: 1.23 to 1.99) had significant heterogeneity (I² = 64%). This can be caused by the use of different samples, including blood, urine, tap water, meconium, and toenails. Our comprehensive findings indicate that As increases the risk of GDM (OR 1.47, 95% CI: 1.11 to 1.95), with considerable heterogeneity (I² = 76%). Although our results of this meta-analysis are consistent with the previous meta-analysis, the effect size has slightly decreased. Additionally, Egger’s test indicated potential publication bias, possibly due to selective reporting and publication preferences. Sensitivity analysis confirmed the robust results, while subgroup analysis explored the sources of heterogeneity. The national subgroups produced different results, possibly due to differences in As intake due to differences in the environment and dietary habits. Metabolic processes and genetic factors may further influence As intake. Subgroup analyses of thresholds revealed statistical significance only for quartile thresholds. The diagnostic criteria, study design, and sample type could not explain the source of heterogeneity. However, the year of publication may provide some explanation for the source of heterogeneity. Future studies could consider the impact of As on GDM by determining diagnostic criteria and dietary habits. While our current analysis of the studies is limited, it is still critical to conduct such studies in the future.

This meta-analysis has many strengths. First, combining data from 13 studies increased the sample size to 2575, improving the accuracy of our findings. Second, the included studies were of medium-to-high quality and were evaluated by the appropriate scales, ensuring highly reliable and credible data. Third, this study conducted a detailed analysis of the subgroups to explore the sources of heterogeneity.

However, this study has several limitations. First, due to differences in diagnostic criteria, test samples, and study designs, there was significant heterogeneity among the studies. Second, the study had publication bias, meaning that there was unpublished literature. Future efforts should integrate these neglected studies to evaluate the overall impact of As on GDM more comprehensively. Third, the use of multiple subgroups results in a smaller sample size, reducing the statistical power and complicating the clarification of effects. Therefore, the heterogeneity among the included studies was not fully explained. Nevertheless, based on existing studies, the current study represents an estimate of the association between As and GDM and provides additional information on different subgroups. However, additional large-scale studies are needed in the future to verify these results.